In: Statistics and Probability
Consider the following regression equation for salaries. Female and Male are dummy variables. The value of Female = 1 for a female and 0 otherwise. The value of Male is 1 for Male and 0 otherwise.
^Sal = 0.88 - 3.58 Female+ 0.25 Age + 1.12 Male
a) Describe in words the classical assumption violated by this equation.
b) What change would you make to the equation to correct the problem mentioned in a)
a) There should be no perfect multicollinearity between the independent variables. This assumption is violated by this equation as there are two dummy variables present in this equation which are depicting exact collinearity as if the person is male then the other will be automatically female, one independent variable is linearity predicted by the other independent variable with the substantial degree of accuracy.. Hence, presence of multicollinearity is there in the equation.
b) Only one dummy variable instead of two i.e. Person. The value of Person is 1 for Male and 0 for Female. This will ensure the absence of Multicollinearity in the equation.